Adding fused_quantized fix

operators/utils.py

@@ -24,7 +24,7 @@ def speak(text):
 
 speak("hello")
 
-MODE = "quantized"
+MODE = "fused_quantized"
 DEVICE = "cuda"
 PROCESSOR = AutoProcessor.from_pretrained("HuggingFaceM4/idefics2-tfrm-compatible")
 BAD_WORDS_IDS = PROCESSOR.tokenizer(
@@ -70,6 +70,17 @@ else:
     raise ValueError("Unknown mode")
 
 
+def reset_awq_cache(model):
+    """
+    Simple method to reset the AWQ fused modules cache
+    """
+    from awq.modules.fused.attn import QuantAttentionFused
+
+    for name, module in model.named_modules():
+        if isinstance(module, QuantAttentionFused):
+            module.start_pos = 0
+
+
 def ask_vlm(image, instruction):
     prompts = [
         "User:",
@@ -87,6 +98,7 @@ def ask_vlm(image, instruction):
     generated_texts = PROCESSOR.batch_decode(generated_ids, skip_special_tokens=True)
 
     text = generated_texts[0].split("\nAssistant: ")[1]
+    reset_awq_cache(model)
     speak(text)
     return text
 

tests/test_idefics2.py

@@ -3,10 +3,16 @@ import torch
 from PIL import Image
 from io import BytesIO
 
-from transformers import AutoProcessor, AutoModelForVision2Seq, AwqConfig
+from transformers import (
+    AutoProcessor,
+    AutoModelForVision2Seq,
+    AwqConfig,
+)
 import awq_ext
 
-MODE = "quantized"
+import time
+
+MODE = "fused_quantized"
 DEVICE = "cuda"
 PROCESSOR = AutoProcessor.from_pretrained("HuggingFaceM4/idefics2-tfrm-compatible")
 BAD_WORDS_IDS = PROCESSOR.tokenizer(
@@ -78,7 +84,19 @@ image1 = download_image(
 )
 
 
+def reset_awq_cache(model):
+    """
+    Simple method to reset the AWQ fused modules cache
+    """
+    from awq.modules.fused.attn import QuantAttentionFused
+
+    for name, module in model.named_modules():
+        if isinstance(module, QuantAttentionFused):
+            module.start_pos = 0
+
+
 def ask_vlm(image, instruction):
+    global model
     prompts = [
         "User:",
         image,
@@ -93,17 +111,20 @@ def ask_vlm(image, instruction):
         max_new_tokens=100,
     )
     generated_texts = PROCESSOR.batch_decode(generated_ids, skip_special_tokens=True)
+    reset_awq_cache(model)
     return generated_texts
 
 
-import time
-
 now = time.time()
+
 print(ask_vlm(image1, "What is this?")[0].split("\nAssistant: ")[1])
 
 print("resp:", time.time() - now)
 import time
 
+
 now = time.time()
 
 print(ask_vlm(image1, "What is this?")[0].split("\nAssistant: ")[1])
+
+print("resp:", time.time() - now)